brod attributes the real monthly bill to topics and teams, finds the waste, and turns every finding into a euro-quantified fix you merge as a pull request — without ever giving any tool write access to your cluster.
curl -fsSL brod.sh | sh
copy
Most clusters bleed money on dead topics, retention overkill, missing compression and over-partitioning. brod ranks every leak in euros — estimate vs measured, always labeled.
Write-capable credentials are rejected at onboarding. brod cannot modify a cluster — by architecture. Fixes ship as diffs you review.
Topic configs, sizes, offsets, consumer-group state. Never message payloads. There is no code path that fetches your data.
Every finding carries a € figure, the exact config change, and an honest basis label. Savings are booked only once a fix is measured.
The open-source CLI gives you a one-shot, euro-ranked scan today. The hosted platform — on the roadmap, not shipped yet — turns that into continuous monitoring, per-team attribution, and fixes delivered as pull requests. Everything below inherits the same three pillars: read-only, metadata only, EU-hosted.
--json / --html; no telemetry by defaultMetadata snapshots every 15 minutes build trend history, so waste is caught as it grows — not once a quarter.
Allocate the real bill to topics and teams with regex/prefix mapping. Showback views and CSV export for finance.
Each finding becomes a GitHub, GitLab, Terraform or Strimzi PR with the diff and evidence. You review and merge — brod never applies.
Euros are booked only after a fix is measured across three snapshots — auditable before/after, never a predicted promise.
Confluent Cloud and AWS MSK agentless, or self-hosted via an outbound-only, read-only collector. No inbound ports.
A plain, forwardable email: new waste in euros, fixes applied, saved-to-date, and the top open recommendations.
Run the CLI today. Get on the list for continuous, per-team cost attribution.